LNK306 buck, non iso, would like more current output
I am designing a buck converter modeled similar to Figure 5, pg 4 of LNK302/304-306 data sheet; with an input stage like AN37 table 9 >1W (Full wave). Universal input levels, 11.25Vout. I've use PIXlsDesigner 7 to derive the passive parts, as well as the app notes etc to make sure they agree. I've also placed in the design the opto feedback, basically to experiment and evaluate both feedback topologies (one at time certainly). The application doesn't require isolation, but would consider doing it to increase current output. The custom tranformer scares me off a little for cost and availability however. I have seen a design where someone used a TNY268 essentially in place of the LNK30x. I'd like to know if I can use other parts from the numerous PI families to get more than .36A out and still use the basic buck converter architecture I have now. Or, must I then utilize the flyback topology with say an EE13 core with the other PI IC's (like PIXlsDesigner steers you to when you select other part families). I only need to get to about .5A at 5V, but might bump the voltage up to linear regulate directly at my mcu's to further clean up.
Thank you, Steven
Comments
Thank you. As far as effeciency is concerned, my max current draw wouldn't be continuous.
I see on the TNY data sheet how they utilize the opto feedback for the opposite polarity.
Your myth buster is making me rethink my strategy.
I don't have analog and digital separation in this application, but I am switching relays that I'm trying to decide if I use higher voltage coils, or all 5V devices. That and a pile of LEDs will draw all the current.
Thanks again for the advice.
SWJ
Just wanted to follow up to give you some feedback. I decided to make the leap and use the TNY274GN in a flyback topology. I used Desg.Exp and Wurth E for proto of the transformer. It all worked excellent first time when boards and parts came in. My only hickup was the inrush current causing my fusible resistor to open after repeated power cycles. I saw buried in an app note about this happening with metal film. I've since solved it. I did add the additional LC's on the output to somewhat separate my digital and relay power; and did not use any linear device to further regulate. You are right, it is a myth, and I'm glad I tried it.
I'm pretty pleased with the product and design tools you have provided.
Thank you,
SWJ
1) Yes you can use the TNY family for more current in a buck converter.
a) However it will be less efficient than a flyback
b) The EN pin of the TNY family has inverted logic compared to the LNK30x family, so your feedback circuit gain needs to have the opposite sign
2) It is a myth to need a linear regulator stage to "clean up" an SMPS output. All you need to do is add more differential filtering to the output to reduce its ripple, and make sure that the SMPS low frequency gain is sufficient to reject 120 Hz ripple. And, a linear regulator will not clean up common mode noise on the output.
2 a) If you have say, two 5V rails for your circuit, where one is "digital" and the other is "analog", the way to prevent cross-talk is to have 2 sets of LC post filters, one for each rail.